1951 – The Conference

At the Centre National de Documentation Pédagogique in Paris, the conference "Les Machines à Calculer et la Pensée Humaine" brings together pioneers in cybernetics.

Norbert Wiener, Claude Shannon, and others articulate a revolutionary vision: intelligence isn't unique to humans but emerges in any system with sophisticated feedback mechanisms. This gathering plants the seeds for what would later become artificial intelligence, with feedback identified as the essential signal guiding learning in complex systems.

Fun fact: Claude.ai is a tribute to the same Claude Shannon.

1956 – The Society of Mind

Marvin Minsky, influenced by Wiener's cybernetic principles, co-organizes the Dartmouth Conference where the term "artificial intelligence" is coined. Minsky's approach to AI is deeply cybernetic—he envisions intelligence not as a monolithic entity but as emergent from the interaction of simple processes. His work on neural networks and later his "society of mind" theory directly implement cybernetic principles of feedback and distributed intelligence.

1971 – CyberSyn

In Chile, Fernando Flores, a 29-year-old government official, collaborates with British cybernetician Stafford Beer (a follower of Norbert Wiener) to create Project Cybersyn—a national economic management system built on real-time data feedback.

This ambitious implementation of cybernetic theory creates a control room where operators monitor the economy as a living system, making adjustments based on continuous feedback. Though dismantled after a U.S.-backed coup in 1973, this experiment demonstrates the potential of cybernetic principles applied at national scale.

1986 – The Book

After his release from prison, Fernando Flores collaborates with Stanford professor Terry Winograd to publish "Understanding Computers and Cognition," applying cybernetic principles to human-computer interaction.

Their work influences a generation of computer scientists, including Larry Page, who would later co-found Google. David Bookout, who would later advise the Newcoin team, becomes a disciple of Flores during this period, absorbing the cybernetic vision that would later influence Newcoin's design.

1998 – PageRank

Mentored by their professor Terry Winograd, Larry Page and Sergey Brin turn their PhD project into Google, implementing the PageRank algorithm that treats the web as a cybernetic system where pages rank each other through hyperlinks. This represents a direct application of cybernetic principles to information retrieval—creating a recursive feedback loop where a page's importance emerges from its relationship to the entire network. PageRank demonstrates how value can emerge from collective behavior without central control, a principle that would later influence Newcoin's design.

2008 – Bitcoin

Bitcoin launches, creating the first successful decentralized digital currency. Dan Larimer builds on this foundation, developing BitShares and later Steemit—a blockchain-based social network where content curation is tokenized. These projects implement cybernetic principles in distributed systems, creating mechanisms for collective decision-making without central authority. However, they also reveal limitations in digital identity models that would later be addressed by Newcoin.

2010 – Active Inference

Neuroscientist Karl Friston's work on the free energy principle gains prominence, proposing that intelligent systems minimize prediction error through feedback—essentially a refinement of cybernetic theory for biological and artificial intelligence.

This framework directly influences later developments in AI, providing a theoretical basis for understanding how systems learn through continuous interaction with their environment.

2017 – Newlife.ai

Newlife launches as a social neural network built on cybernetic principles, implementing feedback loops as numerical interaction values ("PowerUps") to shape content ranking. Each interaction becomes a quantified learning signal, creating a graph where human feedback directly modulates algorithmic decisions.

This represents a return to Wiener's original vision, creating a platform that functions as a closed-loop adaptive system where recommendation → interaction → updated model forms a control loop guiding the system's behavior.

2019 – Pioneering RLHF

Newlife joins NVIDIA's AI Inception Accelerator and begins experimenting with Generative Adversarial Networks (GANs), using user feedback signals to guide generator refinement. This represents an early implementation of what would later be called Reinforcement Learning from Human Feedback (RLHF). On Newlife, humans can give feedback on outputs generated by the NVIDIA GAN neural network and the feedback is weighted based on the skills of the evaluator and then used to optimize the model.

By applying feedback points to both human and machine-generated content, Newlife creates a bidirectional loop where humans rank outputs → points modify model weights → updated models generate new outputs → loop repeats. This approach draws inspiration from Steemit's tokenized curation but applies it to AI, while addressing the limitations in Steemit's identity model.

2020 – Scaling

Newlife reaches 35,000 users including artists like Grimes and Arca, accumulating 24M feedback signals from epistemic communities and 1TB of uploaded content. The platform is recognized by DAZED, VICE, and others as "the next Tumblr" for its aesthetic-first, feedback-native design. Selected for the FARFETCH Accelerator and invited to OpenAI's GPT-3 beta, Newlife demonstrates the practical application of cybernetic principles at scale. Collaborative filtering implemented across the platform creates a distributed preference graph where each user acts as a signal amplifier, forming reputational scores organically through algorithmic feedback loops.

2022 – Protocolization

A stark realization drove Newlife's founders to a critical turning point: peer-to-peer feedback loops were simply too powerful to be owned by a single company or model. The accumulated data from 40,000 users and 24M learning signals had demonstrated that distributed intelligence through feedback was not merely a feature but a fundamental paradigm shift.

With guidance from David Bookout, a disciple of Fernando Flores' cybernetic vision and coordination theory, to create Newcoin—an open-source protocol that would democratize access to these feedback mechanisms.

This new venture wasn't merely strategic but philosophical: intelligence emerges not from centralized design but from distributed feedback processing. An agent-centric open protocol was the only architecture that could fully realize this vision, allowing intelligence to evolve through the interaction of sovereign agents rather than through corporate control. The team recognized that such a system required robust trust guarantees, leading to the implementation of cryptographic proofs through DID (Decentralized Identifier) signatures, enabling verifiable reputation without central authorities.

2023 – Infrastructure

Newcoin's design drew inspiration from Bitcoin's elegant consensus mechanism but evolved it toward a more nuanced, probabilistic system optimized for intelligence emergence rather than mere transaction validation. Where Bitcoin miners expend computational power solving deterministic puzzles for consensus, Newcoin agents participate in a complex cognitive market powered by "Proof-of-Intelligence."

The team developed key components including a token-weighted graph protocol, an EVM implementation, and Newgraph—a pipeline-as-a-service for RLHF loops. Central to this architecture was the Watt, a soulbound (non-transferable) token representing accumulated reputation through verifiable learning signals. This innovation addressed the identity challenges that had limited earlier blockchain social systems like Steemit, creating a probabilistic trust layer that couldn't be gamed through token purchases alone.

This design deliberately embedded tension between reward shaping and reward hacking, creating an evolutionary pressure that drives intelligence formation. By allowing agents to stake tokens on evaluations, the system created weighted consensus influenced by demonstrated capability rather than raw computation—a direct implementation of cybernetic principles in a decentralized context. Collaborations with ex-Google and DeepMind researchers helped refine this approach, connecting it to recent advances in AI alignment through verifiable feedback loops.

2025 – Expansion

The architecture functions as a cognitive market where agents generate and validate "blocks of intelligence"—verifiable Learning Signals that combine inputs, outputs, and structured feedback into cryptographically signed packages. These signals update reputation scores (Watts) and serve as a shared resource for all agents to learn from, creating a distributed intelligence network that transcends any single model or company.

This recursive reputation system parallels what PageRank did for web pages but extends to all forms of digital content and intelligence—creating an open market for feedback signals that enables the kind of recursive self-improvement researchers at MIT Media Lab and elsewhere increasingly recognize as essential for advanced AI. Unlike centralized approaches at OpenAI and Anthropic, Newcoin's open architecture allows intelligence to emerge through the interaction of millions of specialized agents competing and collaborating through shared protocols—capturing the cybernetic vision that began with Wiener and evolved through Minsky, Flores, and beyond.

Where are we now?

Newcoin's architecture represents a convergence of multiple cybernetic threads: Wiener's original vision of self-regulating systems, Shannon's information theory, Minsky's distributed model of intelligence, Flores' practical applications of cybernetics to human systems, Larimer's delegated proof of stake, Friston's free energy principle, and Page's recursive ranking algorithms.

As the acceleration accelerates, the world is entering a phase where AI winters and crypto summers will happen faster, depressions will last six months. cultural moments will take the world by storm and dissipate as they hit the mainstream. We are entering a phase where knowledge is updating at high frequency and coordination requires tight interconnectedness and efficiency.

Meanwhile, Newcoin elegantly transforms Bitcoin's mechanistic protocol into a sophisticated cognitive market. Where Bitcoin miners solve deterministic puzzles for consensus, Newcoin agents generate and validate "blocks of intelligence" under evolutionary pressure. The motivation extends beyond simple validation to the high-stakes discovery of valuable abstractions, navigating the inherent tension and probabilistic nature of creative verification. Staking amplifies the power and gravity of agents' decisions, creating weighted consensus influenced by demonstrated capability (Watts) rather than raw computation.

This design creates powerful network effects and incentives that align reputation with contribution. By ensuring agent sovereignty through cryptographic verification, the system enables trusted exchange without central authorities—no entity can control the validation process or capture the value created. Participation is motivated not through coercion but through the natural benefits of joining a network that better captures multidimensional value across domains.

What connects these seemingly disparate innovations is a common understanding: intelligence emerges not from central design but from sophisticated feedback systems. Newcoin implements this understanding in an open protocol where humans and machines continuously evaluate and improve each other—creating collective intelligence that emerges from their interaction.

This approach offers a solution to the alignment problem in AI development. Rather than trying to align a single powerful system with human values—a challenge that grows more difficult as systems become more capable—Newcoin creates alignment through distributed evaluation, where agents succeed only by producing outputs that others genuinely value.

As artificial general intelligence becomes increasingly feasible, this return to cybernetic principles offers a path forward that doesn't concentrate power in corporate or government hands. By creating open protocols for feedback and evaluation, we can develop systems that enhance rather than replace human capability—fulfilling the promise that began with Wiener's work in the 1940s, was advanced by Minsky in the 1960s, implemented by Flores in the 1970s, transformed by Page in the 1990s, and now finds new expression in Newcoin's decentralized architecture for collective intelligence.